I just finished my exams so that’s why I haven’t been doing much lately. Today, I realised that my tic tac toe neural network powered genetic algorithm thing isn’t going to work the way it is at currently, but I did learn a few things.
I picked up the code that I had written a few weeks ago and I realise that it’s a pretty convoluted way of making tic tac toe and even more so if I wanted to create a neural network. First of all the input are not numbers, I made them letters (“x” and “y”) and blanks spaces. This would have made it hard to put as input for the neural network. If I were to change this, all my formatting for printing the game would have to change, and to be honest, I’ve forgotten half of how I coded that part, so that would have been a pain to do.
I did learn about this function that returns the index of the max of a list (so you could say that it wasn’t all for nothing). Also http://stats.stackexchange.com/questions/181/how-to-choose-the-number-of-hidden-layers-and-nodes-in-a-feedforward-neural-netw was quite informative.
I haven’t given up on this, I just to be honest, cannot be bothered rewriting the whole thing right now. I’ll do it maybe, when I’m bored and have forgotten my current frustration.
- If and when I do redo everything, here is what I need to make sure to do (ie, a note to self):
- create an easy way to reference the position of each square (ideally use one list instead of 3 in one for the squares)
- don’t use strings as the pieces, use numbers and rewrite the print function to replace them with letters.
- actually write out what I plan to do, so that I don’t make mistakes like this again (ie, not planning for what I plan to do.
- Over the exam week(s), I happened across Game Theory and /r/GAMETHEORY. And so I took a look into it. It’s pretty interesting stuff, and I’m thinking about “restructuring” this blog to make it more general. Sort of like a general blog where I can post my thoughts/projects relating to ideas that are interesting to me.